A. SPECIFIC AIMS Mental retardation (MR) is one of the most common neurodevelopmental conditions and presents as developmental delay in young children. However, not all children who initially present with developmental delay have MR. Indeed, some children with developmental delay catch up with their peers, in an atypical course to normal development (AC). In a child with developmental delay who is less than 3 years of age, it is difficult to distinguish between MR and AC when the results of conventional clinical investigations are normal. From a clinical perspective, it is important to distinguish between these two conditions reliably as the treatment and prognosis are different. We recently performed a diffusion tensor imaging (DTI) MRI study in children with nonautistic developmental delay whose routine investigations (MRI, genetic testing, etc.) were normal and found on tractography that the arcuate fasciculus (a tract involved in language function) was unidentifiable/underdeveloped in approximately half of the children studied (arc- group), compared to typically developing children (TD) in whom it was always clearly identifiable (Sundaram et al., 2008). Since such a pronounced abnormality in brain white matter structure is unlikely to result in the relatively rapid developmental improvement seen in AC, we hypothesize that children with underdeveloped/unidentifiable arcuate fasciculus (arc- endophenotype) will fall into the MR rather than into the AC group. We also hypothesized that in children with idiopathic developmental delay (no doubt a heterogeneous group), the use of DTI tractography may provide a new strategy to elucidate potential etiologies. In this application, we propose to use probabilistic tractography to quantify the strength of the arcuate fasciculus, and we hypothesize that this technique can be used as a biomarker to designate children with developmental delay who are unlikely to make full recovery, thus allowing early diagnosis. We will also use the automated and objective Tract Based Spatial Statistics (TBSS) software to analyze the properties of white matter (FA, ADC, etc) tracts of the whole brain in these children. We will perform TBSS and tractography in developmentally delayed children with and without an identifiable arcuate fasciculus and in typically developing (TD) children. We further hypothesize that the arc- endophenotype may facilitate the discovery of further genetic causes of MR beyond what is currently known. A major advantage and innovation of using this DTI endophenotype over the more general clinical phenotype of developmental delay is reduced heterogeneity and hence increased ability to identify a genetic cause. Since proper axon path-finding is a necessary pre-requisite for the normal development of white matter tracts, we hypothesize that mutations in the genes of axon guidance network may be responsible for the arc- endophenotype. Thus, the overall purpose of the present study is to determine whether the arc- endophenotype measured by DTI is an imaging biomarker for MR and a useful endophenotype for the discovery of genetic mutations causing MR. Specific Aim 1: To determine whether diffusion tensor imaging (DTI) is an imaging biomarker for the discrimination of idiopathic mental retardation (MR) from atypical course to normal development (AC) in children between 18 - 36 months of age who present with delayed developmental milestones and fail to show a specific etiology after standard clinical evaluation. Hypothesis 1A. Developmentally delayed children with underdeveloped/unidentifiable arcuate fasciculus (based on deterministic tractography) (arc- endophenotype) will have a lower probability of connection (based on probabilistic tractography) between receptive language (inferior parietal, superior and middle temporal) and expressive language (middle and inferior frontal) areas as compared to developmentally delayed children with identifiable arcuate fasciculus (arc+ endophenotype) and typically developing (TD) children. In addition, expressive and receptive language areas will show significantly lower FA using TBSS in the arc- group as compared to arc+ group and TD group. Hypothesis 1B. Children with developmental delay who have been identified with a low probability of connection between receptive and expressive language areas will fall primarily into the MR group rather than into the AC group. Hypothesis 1C. The probability of connection between receptive and expressive language areas as assessed using probabilistic fiber tracking will positively correlate with language scores obtained from neuropsychological evaluation. Specific Aim 2: To determine the genetic basis of arc- endophenotype in children with developmental delay. Hypothesis 2A: Mutations will be found in genes along the axon guidance pathway in developmentally delayed children with arc- endophenotype. Hypothesis 2B: Genomic copy number variation will be found in developmentally delayed children with arcendophenotype.